Guru
- Posts: 3647
- Joined: 20 Oct 2007, 11:47
- Location: Nottinghamshire: The Home of Robin Hood..... UK
06 Jan 2012, 19:53
I recently had an interesting chat with an electrical engineer employed by Frontier Semiconductors from '72-'76 during the heyday of LED watches.
I asked him a question regarding the tendency for some of the modules to fail due to battery acid damaging the film resistor on the front of the module, he granted me permission to share his reply with the forum.......
"The thick film resistors that are printed on the ceramic substrate wasn't a problem for us during the '70's while I was at Frontier. Of course, we didn't anticipate what would happen 35+ years down the road. I've seen several references by others in the various forums I've visited mentioning problems with the resistors. Might very well be, as thick film resistors normally use resistance metal oxides with glass-based materials to form a paste for printing on the substrate. After burning off the solvents and after few other heat treatments the glass filled resistive material has a sponge like structure with random inclusions and defects. These films will absorb gases and moisture (if not hermetically sealed) causing drifts and long term instability. I do think the problem might be caused from the "trimming" of the resistance value by the manufacturing company that was doing the die mount and wire bonding for us in the Philippines. When the resistors are printed onto the substrate, and before trimming, the resistors are sealed by firing at temperature. At that point they should have been protected from moisture, gasses etc. Prior to shipping the finished substrate to Frontier, they would trim the resistor value, using a LASER, not a DIAMOND SAW as one post I read, to the value Frontier had specified. Once that Laser did it thing, that persisted layer seal was broken and would allow moisture etc. to be absorbed. I've looked at several modules I have here on the bench and can now see that we made a big mistake by not requiring the manufacture to reseal after trimming.
I would think that those resistors could be replaced with garden variety resistors that one could purchase through various electronic distributors and bring a dead module back to life. Both of those resistors are connected into the quartz oscillator circuit and aren't real critical of value. The "long" resistor that is printed on the substrate is?the one I would highly suspect if the oscillator doesn't run. It is hooked up directly in
parallel with the quartz crystal and had a nominal trimmed value of about 15Meg Ohms. If one were going to replace it with a store bought resistor, anything between 10Meg Ohms to 20 Meg Ohms should work just fine. What you would want to do is try and break the continuity of the printed resistor on the substrate by gently cutting or scraping, maybe with a small rotary tool, so that it would no longer have any effect on the oscillator circuit. You could then solder a small 1/8 watt resistor to the same silver pads on the substrate that the quartz crystal leads are attached too. The smaller resistor was inserted into the oscillator circuit to limit the drive current to the quartz crystal so that the watch chip in a non-display mode would consume no more than 3uA (3 millionths amp). Its value was about 150K - 180K Ohms, another standard value that one could purchase off the shelf. Normally thick film resistors go down in value when they absorb moisture, gasses etc. The small resistor could go down in value to almost nothing and the oscillator should still run OK, so I wouldn't suspect it, even if it had changed value. The long resistor will definitely stop the oscillator from running if it's value changes drastically. Increasing in value to beyond 20 - 30 Meg Ohms or decreasing to a low value due to the absorption of gasses or moisture will?stop the oscillator from ever starting. So the bottom line is to get both a 150 -180K Ohm, and 10 - 15Meg Ohm 1/8watt resistors to replace the thick film resistors."
Follow up reply....
"I didn't realize or think about the resistor issue until you brought it up. The light went on and I got to looking at some substrates here on the bench. It was obvious that after the Laser trim of the resistors, there was no sealing the exposed edges of the Laser cut. DUH!!! Wow, that's a real screwup on Frontiers part. The replacement resistors for the thick film could also be surface mount, if you can get them in the high ohmic range of 10 - 15Meg Ohm. The 150 - 180K Ohm shouldn't be to hard to find in surface mount style. 1/8th or even 1/10th watt resistors are pretty small and should fit into the case just fine. They also have leads for soldering whereas surfuce mount doesn't, just metalized end caps to solder too.
If it is of any interest to you, here's a link to the US Patent issued to myself and others for the Frontier Alarm Watch. There are 4 individuals named on that patent, John Fassett,???Ivan Rondas, Claus Sondermann and myself. John Fassett was the Division Head, Ivan Rondas was the Department Head and Claus and I lowly working engineers. John and Ivan had zip, ziltch, zero to do with the design of that watch, but being honchos they had their names added to the patent."
http://www.patentgenius.com/patent/4068461.html
.
I asked him a question regarding the tendency for some of the modules to fail due to battery acid damaging the film resistor on the front of the module, he granted me permission to share his reply with the forum.......
"The thick film resistors that are printed on the ceramic substrate wasn't a problem for us during the '70's while I was at Frontier. Of course, we didn't anticipate what would happen 35+ years down the road. I've seen several references by others in the various forums I've visited mentioning problems with the resistors. Might very well be, as thick film resistors normally use resistance metal oxides with glass-based materials to form a paste for printing on the substrate. After burning off the solvents and after few other heat treatments the glass filled resistive material has a sponge like structure with random inclusions and defects. These films will absorb gases and moisture (if not hermetically sealed) causing drifts and long term instability. I do think the problem might be caused from the "trimming" of the resistance value by the manufacturing company that was doing the die mount and wire bonding for us in the Philippines. When the resistors are printed onto the substrate, and before trimming, the resistors are sealed by firing at temperature. At that point they should have been protected from moisture, gasses etc. Prior to shipping the finished substrate to Frontier, they would trim the resistor value, using a LASER, not a DIAMOND SAW as one post I read, to the value Frontier had specified. Once that Laser did it thing, that persisted layer seal was broken and would allow moisture etc. to be absorbed. I've looked at several modules I have here on the bench and can now see that we made a big mistake by not requiring the manufacture to reseal after trimming.
I would think that those resistors could be replaced with garden variety resistors that one could purchase through various electronic distributors and bring a dead module back to life. Both of those resistors are connected into the quartz oscillator circuit and aren't real critical of value. The "long" resistor that is printed on the substrate is?the one I would highly suspect if the oscillator doesn't run. It is hooked up directly in
parallel with the quartz crystal and had a nominal trimmed value of about 15Meg Ohms. If one were going to replace it with a store bought resistor, anything between 10Meg Ohms to 20 Meg Ohms should work just fine. What you would want to do is try and break the continuity of the printed resistor on the substrate by gently cutting or scraping, maybe with a small rotary tool, so that it would no longer have any effect on the oscillator circuit. You could then solder a small 1/8 watt resistor to the same silver pads on the substrate that the quartz crystal leads are attached too. The smaller resistor was inserted into the oscillator circuit to limit the drive current to the quartz crystal so that the watch chip in a non-display mode would consume no more than 3uA (3 millionths amp). Its value was about 150K - 180K Ohms, another standard value that one could purchase off the shelf. Normally thick film resistors go down in value when they absorb moisture, gasses etc. The small resistor could go down in value to almost nothing and the oscillator should still run OK, so I wouldn't suspect it, even if it had changed value. The long resistor will definitely stop the oscillator from running if it's value changes drastically. Increasing in value to beyond 20 - 30 Meg Ohms or decreasing to a low value due to the absorption of gasses or moisture will?stop the oscillator from ever starting. So the bottom line is to get both a 150 -180K Ohm, and 10 - 15Meg Ohm 1/8watt resistors to replace the thick film resistors."
Follow up reply....
"I didn't realize or think about the resistor issue until you brought it up. The light went on and I got to looking at some substrates here on the bench. It was obvious that after the Laser trim of the resistors, there was no sealing the exposed edges of the Laser cut. DUH!!! Wow, that's a real screwup on Frontiers part. The replacement resistors for the thick film could also be surface mount, if you can get them in the high ohmic range of 10 - 15Meg Ohm. The 150 - 180K Ohm shouldn't be to hard to find in surface mount style. 1/8th or even 1/10th watt resistors are pretty small and should fit into the case just fine. They also have leads for soldering whereas surfuce mount doesn't, just metalized end caps to solder too.
If it is of any interest to you, here's a link to the US Patent issued to myself and others for the Frontier Alarm Watch. There are 4 individuals named on that patent, John Fassett,???Ivan Rondas, Claus Sondermann and myself. John Fassett was the Division Head, Ivan Rondas was the Department Head and Claus and I lowly working engineers. John and Ivan had zip, ziltch, zero to do with the design of that watch, but being honchos they had their names added to the patent."
http://www.patentgenius.com/patent/4068461.html
.